TM 665 Project Planning _amp; Control Dr. Frank Joseph Matejcik

Document Sample
TM 665 Project Planning _amp; Control Dr. Frank Joseph Matejcik Powered By Docstoc
					       TM 745 Forecasting for
       Business & Technology
      Dr. Frank Joseph Matejcik
    10th Session 4/19/10:
    Chapter 10 Forecast Implementation




South Dakota School of Mines and
     Technology, Rapid City
         Tentative Schedule
         Chapters     Assigned          Chapters Assigned
   25-Jan 1    problems 1,4,8      5-Apr Easter
                e-mail, contact   12-Apr 8 Problem 6
    1-Feb 2    problems 4,8,9     19-Apr 10(6th) 9(5th)
    8-Feb 3    problems 1,5,8,11
   15-Feb      President’s Day    26-Apr EIPI(QTS) Or
   22-Feb 4    problems 6, 10           9 Data Mining?
    1-Mar 5    problems 5, 8      3-May Final
    8-Mar      Break
   15-Mar      Exam 1 Ch 1-4 Revised
   22-Mar 6    problems 4, 7
   29-Mar 7    3,4,5(series A) 7B
Frank Matejcik SD School of Mines & Technology   2
         Web Resources
           Class Web site on the HPCnet system
           http://sdmines.sdsmt.edu/sdsmt/d
            irectory/courses/2010sp/tm745M0
            01
           Answers will be online. Linked from ^
           I have gotten D2L and Elluminate!
            sites, and have gotten started on
            Elluminate! documentation.

Frank Matejcik SD School of Mines & Technology   3
         Web Resources
           Class Web site on the HPCnet system
           http://sdmines.sdsmt.edu/sdsmt/direct
            ory/courses/2008sp/tm745M021
           Streaming video
            http://its.sdsmt.edu/Distance/
           Answers will be online. Linked from ^
           The same class session that is on the DVD is
            on the stream in lower quality.
            http://www.flashget.com/ will allow you to
            capture the stream more readily and review
            the lecture, anywhere you can get your
            computer to run.

Frank Matejcik SD School of Mines & Technology   4
         Agenda & New Assignment
        Chapter 10(6th) 9(5th) no problems
        Final is in two weeks
        Study guide isn’t posted, yet
        Chapter 10(6th) 9(5th) Forecast
         Implementation




Frank Matejcik SD School of Mines & Technology   5
      10(6th) 9(5th) Forecast
      Implementation

           Keys (a list)
           Forecast Process (steps)
           Choosing the right forecast
           New Product
           Artificial Intelligence



Frank Matejcik SD School of Mines & Technology   6
         Keys to Obtaining Better
         Forecasts
   1. Understand what forecasting is & is not
       Focus on management processes & controls, not
        computers; Establish forecasting group
       Implement management control systems before
        selecting forecasting software
       Derive plans from forecasts
       Distinguish between forecasts and goals
       Forecasting is acknowledged as a critical
       Accuracy emphasized; not game-playing
Frank Matejcik SD School of Mines & Technology   7
        Keys to Obtaining Better
        Forecasts
           2. Forecast demand, plan supply
               Don’t use shipments as actual demand
               Identify sources of demand information
               Build systems to capture key demand data
               Get improved customer service &
                capital planning



Frank Matejcik SD School of Mines & Technology   8
           Keys to Obtaining Better
           Forecasts
      3. Communicate, cooperate, & collaborate
          Avoids duplication & Mistrust of "official“
           forecast
          Creates understanding of impact throughout
          Establish a cross-functional approach to
           forecasting




Frank Matejcik SD School of Mines & Technology   9
           Keys to Obtaining Better
           Forecasts
      3. Communicate, cooperate, & collaborate
          Establish an independent forecast group that
           sponsors cross-functional collaboration
          All relevant information used to generate
           forecasts
          Forecasts trusted by users
          More accurate & relevant forecasts


Frank Matejcik SD School of Mines & Technology 10
         Keys to Obtaining Better
         Forecasts
   4. Eliminate islands of analysis
       Mistrust & inadequate information leading
        different users to create their own forecasts
       Build 1 "forecasting infrastructure"
       More accurate, relevant, & credible forecasts
       Provide training for both users
        & developers of forecasts
       Optimized investments in information &
        communication systems
Frank Matejcik SD School of Mines & Technology 11
          Keys to Obtaining Better
          Forecasts
     5. Use tools wisely
         Relying solely on qualitative or quantitative
         Integrate quantitative & qualitative methods
         Identify sources of improved accuracy &
          increased error
         Provide instruction
         Process improvement in efficiency &
          effectiveness

Frank Matejcik SD School of Mines & Technology 12
         Keys to Obtaining Better
         Forecasts
    6. Make it important
        Have accountability for poor forecasts
        So developers can understand forecast uses
        Training developers to understand implications
         of poor forecasts
        Include forecast performance in
         performance plans & reward systems
        Striving for accuracy & credibility

Frank Matejcik SD School of Mines & Technology 13
          Keys to Obtaining Better
          Forecasts
     7. Measure, measure, measure
         Know if the firm is getting better
         Measure accuracy at relevant levels of
          aggregation
         Ability to isolate sources of forecast error
         Establish multidimensional metrics
         Incorporate multilevel measures
         Measure accuracy whenever &
          wherever forecasts are adjusted
Frank Matejcik SD School of Mines & Technology 14
       Keys to Obtaining Better
       Forecasts
           7. Measure, measure, measure
               Forecast performance can be included in
                individual performance plans
               Sources of errors can be isolated and
                targeted for improvement
               Achieve greater confidence in
                forecast process


Frank Matejcik SD School of Mines & Technology 15
       The Forecast Process
           1. Specify objectives
               Articulate role of forecast in decisions
               If forecasts don’t effect decisions, Why?
           2. Determine what to forecast
               Sales: revenue or units?
               weekly, annually, quarterly?
               Communicate with user

Frank Matejcik SD School of Mines & Technology 16
       The Forecast Process
           3. Identify time dimensions
               Horizon
               Frequency
               Urgency
           4. Data considerations
               Internal needs database
                management & disaggregation:
                time, unit, region
               External gov’t, trade association
Frank Matejcik SD School of Mines & Technology 17
       The Forecast Process
           5. Model selection (next section)
           6. Model evaluation
               Less important for subjective methods
               Use holdout method if quantitative
               Go back to step five if a problem
           7. Forecast preparation
               Try for multiple & multiple types

Frank Matejcik SD School of Mines & Technology 18
          The Forecast Process
         8. Forecast presentation
             Management must understand & be
              confident (corporate culture)
             Oral & written
                  same time & same level
                  be generous with charts etc.
         9. Tracking results
             process continues
             reviews open, objective, & positive
Frank Matejcik SD School of Mines & Technology 19
        Choosing the Right
        Forecasting Techniques
       Few hard and fast rules (guidelines)
       Focus on data, time, & personnel
       Subjective Methods
           Sales force composite
                short to medium term
                Preparation time is quick once set up
           Customer surveys
                medium to long term, take 2-3 months
                survey research is a profession
Frank Matejcik SD School of Mines & Technology 20
        Choosing the Right
        Forecasting Techniques
       Subjective Methods
           Jury of Executive Opinion
                Requires Expertise
                Is relatively quick to prepare
           Delphi
                long to medium term
                useful for new products
                can be slow; computers help
                alternatives are better

Frank Matejcik SD School of Mines & Technology 21
        Choosing the Right
        Forecasting Techniques
       Objective Methods
           Naive (little data, sometimes good)
           Moving Averages (easy, little data)
           Exponential Smoothing Simple
                Need to establish weight
                Easy to compute, quick
           Adaptive response ES
                short term, no seasonality
                Users need little background
Frank Matejcik SD School of Mines & Technology 22
        Choosing the Right
        Forecasting Techniques
       Objective Methods
           Holt's ES
                short term, no seasonality, trend included
                Users need little background
           Winters’ ES
                short term, seasonality, trend included
                Need 4 or 5 observations per season
                Need computer for updates
                Users need little background
                 (tell them about weighted dates)
Frank Matejcik SD School of Mines & Technology 23
          Choosing the Right
          Forecasting Techniques
     Objective Methods
         Regression-Based
             Trend (10 observations, quick to develop, easy for
              users, modest developer skills)
             Trend with Seasonality (Need 4 or 5 observations
              per season, short to medium term, need
              a computer, usually little sophistication)
             Causal (10 observations per independent
              variable, short, medium, or long term,
              developers need regression skills.)

Frank Matejcik SD School of Mines & Technology 24
         Choosing the Right
         Forecasting Techniques
   Objective Methods
       Time-Series Decomposition (two peaks and two
        troughs per cycle, 4 to 5 seasons of data, can
        use turning points, short to medium range,
        modest sophistication, managers like it.)
       ARIMA (managers don’t like it, it takes
        a skillful developer, Need a computer
        to do ACF and PACF plots)

Frank Matejcik SD School of Mines & Technology 25
       New Product Forecasting
           Product Life Cycle (PLC) curve




Frank Matejcik SD School of Mines & Technology 26
          New Product Forecasting
         Analog forecasts
             Similar products
             Think Christmas movie toys
         Test marketing
             Pick a “smaller” representative place
             Ex. given is Indianapolis
         Product clinics (panel lab study)
         Type of Product Affects NPF
Frank Matejcik SD School of Mines & Technology 27
        Artificial Intelligence and
        Forecasting (5th)
           Expert systems
           Neural Networks

   Data Mining (6th)
      Works with large databases (unrelated?)
      Diapers and Beer
      Sports Cars have fewer insurance claims
Frank Matejcik SD School of Mines & Technology 28
           Summary

          Difficult task; many considerations
          New opportunities




Frank Matejcik SD School of Mines & Technology 29
       Using “ProCastTM” in
       ForecastXTM to Make Forecasts
           It is okay now that you know what you
            are doing.
           You understand that a selection method
            is choosing the best of things that you
            already know.




Frank Matejcik SD School of Mines & Technology 30

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:4
posted:2/13/2012
language:
pages:30